adaptation to increased dietary salt intake in...

Adaptation to Increased Dietary Salt Intake in the RatRole of Endogenous Nitric Oxide

Pamela J. Shultz and Jonathan P. TolinsDepartment of Medicine, Veterans Affairs Medical Center and University of Minnesota School of Medicine,Minneapolis, Minnesota 55417

Abstract

Previous studies have suggested that nitric oxide (NO) plays arole in regulation of renal vascular tone and sodium handling.Wequestioned whether the effects of NOsynthase inhibitionon renal function are direct or due to increased renal perfusionpressure (RPP) and whether stimulation of endogenous NOactivity plays a role in adaptation to increased dietary salt in-take. Intrarenal arterial infusion of the NOsynthase inhibitorNG-monomethyl-L-arginine (L-NMMA) in control rats re-sulted in decreased glomerular filtration rate, renal vasocon-striction, natriuresis, and proteinuria. WhenRPPwas held atbasal levels with a suprarenal aortic snare, L-NMMAhad simi-lar hemodynamic effects but decreased sodium excretion anddid not induce proteinuria. Exposure of rats to high salt intake(1% NaCl drinking water) for 2 wk induced increased serumconcentration and urinary excretion of the NOdecompositionproducts, NO2+ NO3. Urinary NO2+ NO3and sodium excre-tion were significantly correlated. Compared with controls,chronically salt-loaded rats also demonstrated enhanced renalhemodynamic responses to NOsynthase inhibition. Wecon-clude that the endogenous NOsystem directly modulates renalhemodynamics and sodium handling and participates in therenal adaptation to increased dietary salt intake. Enhanced NOsynthesis in response to increased salt intake may facilitatesodium excretion and allow maintenance of normal blood pres-sure. (J. Clin. Invest. 1993. 91:642-650.) Key words: endothe-lium-derived relaxing factor * nitric oxide - NG-monomethyl-L-arginine * nitro-L-arginine-methylester - sodium * natriuresis

Introduction

In 1980, Furchgott and Zawadski ( 1 ) demonstrated that vascu-lar relaxation induced by the muscarinic agent acetylcholinewas dependent on the presence of a functionally intact endothe-lium and postulated the release by endothelial cells of a labile

Parts of this work were presented at the Annual Meetings of the Ameri-can Federation for Clinical Research, Baltimore, MD, 1-4 May 1992,and have been published in abstract form ( 1992. Clin. Res. 40:353A).

Address reprint requests to J. P. Tolins, M. D., Veterans Adminis-tration Medical Center Renal Section 11 -J, 1 Veterans Drive, Minne-apolis, MN55417.

Receivedfor publication 21 May 1992 and in revisedform 3 Sep-tember 1992.

The Journal of Clinical Investigation, Inc.Volume 91, February 1993, 642-650

factor termed endothelium-derived relaxing factor (EDRF).1Since that initial observation our understanding of the EDRFsystem has expanded rapidly. It is currently believed thatEDRFis nitric oxide (NO), which is derived from the terminalguanidino nitrogen atom of the aminoacid, L-arginine (2-6).A soluble enzyme, currently termed NOsynthase, is involvedin the formation of NOand L-citrulhine from L-arginine (7, 8).Synthesis of NOcan be inhibited by L-arginine analogues suchas L-N0-monomethyl-arginine (L-NMMA) and L-N0-nitro-arginine-methylester (L-NAME) (9-1 1 ). In vitro studies havedemonstrated that NOactivates soluble guanylate cyclase intarget cells, such as vascular smooth muscle ( 12) and glomeru-lar mesangial cells ( 13), inducing a rise in levels of guanosine3',5'-cyclic monophosphate (cGMP), which mediates subse-quent physiological effects. Urinary cGMPlevels have alsobeen demonstrated to reflect activity of the NO system invivo ( 14).

Using this understanding of the basic physiology of the en-dogenous NOsystem, attention has recently focused on therole of this mediator in the regulation of vascular tone in vivo.By observing the effects in vivo of endothelium-dependent ago-nists, such as acetylcholine, and NOsynthase inhibitors, suchas L-NMMAand L-NAME, it has become clear that NOplays aprominent role in regulation of renal hemodynamics and ex-cretory function ( 15-20). When infused into the renal arteryof the rat, L-NMMAinduces a fall in GFRand renal vasocon-striction and a rise in filtration fraction ( 15 ). Micropuncturestudies in the rat during intravenous infusion of L-NMMAhave demonstrated glomerular arteriolar vasoconstriction,with a predominant effect on the efferent arteriole, a decreasein the ultrafiltration coefficient, Kf, and an increase in glomer-ular capillary pressure, Pgc ( 16 ).

Several investigators have reported that NOsynthase inhibi-tion in vivo induces natriuresis and diuresis ( 16, 17, 20). Al-though some investigators have suggested that the increasedexcretion of sodium and water during NOinhibition is due to"pressure natriuresis" ( 18, 21 ), others have hypothesized a di-rect tubular action of NO, observing that NOsynthase inhibi-tion results in decreased proximal reabsorption of sodium(20). This issue has been complicated by the difficulty in sepa-rating direct renal effects of NOsynthesis inhibition from sys-temic hemodynamic effects, such as increased arterial pressure

1. Abbreviations used in this paper: cGMP, guanosine 3',5'-cyclicmonophosphate; EDRF, endothelium-derived relaxing factor; L-NAME, nitro-L-arginine-methylester; L-NMMA, N0-monomethyl-L-arginine; NO, nitric oxide; NO2, nitrite; NO3, nitrate; PAH, para-aminohippurate; RPF, renal plasma flow; RPP, renal perfusion pres-sure; RVR, renal vascular resistance.

642 P. J. Shultz and J. P. Tolins

and subsequent renal autoregulation, in the various experimen-tal models. Chen and Sanders (22) recently suggested that activ-ity of the endogenous NOsystem may be related to sensitivityor resistance to salt-induced hypertension in rats, further sug-gesting a role for NOin renal sodium handling, and, in particu-lar, in adaptation to increased dietary salt loads.

Since NOis a very labile substance, direct measurement ofNOhas proven to be difficult, especially in vivo. Thus, much ofthe in vivo work to date has been limited to interpretation ofthe responses to NOsynthase inhibitors. Although this is auseful approach, it is clear that it would be important to de-velop markers of NOactivity that could be quantitated in vivo.Tolins et al. ( 14), previously demonstrated that, in the rat,hemodynamic responses to the endothelium-dependent va-soactive agonist, acetylcholine, correlated with urinary excre-tion of cGMP, and that the cGMPand hemodynamic re-sponses to acetylcholine infusion could be inhibited by adminis-tration of L-NMMA. These investigators suggested that urinarycGMPexcretion could potentially be used as a biologicalmarker of EDRFactivity. More recently, Marletta et al. (23)demonstrated that NOdecomposes rapidly in biological solu-tions into nitrite (NO2) and nitrate (NO3). These stable de-composition products can be measured as markers of NOpro-duction in vitro, as demonstrated in macrophages (23), vascu-lar smooth muscle (24), and glomerular mesangial cells (25) inculture. Recently, NO2 + NO3have been used as markers ofNOactivity in vivo. Shultz and Raij (26) reported increases inlevels of NO2+ NO3in the serum and urine of rats treated withendotoxin. The increase in NO2+ NO3levels was inhibited bycoadministration of the NOsynthase inhibitor, L-NAME. Fur-thermore, this inhibitory effect was overcome by administra-tion of excess L-arginine, confirming that the NO2+ NO3mea-sured in the serum and urine after endotoxin reflected changesin NOproduction. Thus, urinary levels of cGMP, as well asserum and urine levels of NO2+ NO3, may be useful as in vivo"footprints" of activity of the NOpathway.

In the present studies we have used both the pharmacologi-cal inhibitors of NOsynthase and markers of in vivo NOpro-duction to address the following questions: What is the effect ofinhibition of NOsynthase on renal hemodynamics and excre-tory function when the influence of systemic hemodynamicresponses is eliminated? Does the endogenous NOsystem playa role in adaptation of the normal animal to increased dietarysalt intake?

Methods

Studies were performed in male, Sprague-Dawley rats (275-375 g;Harlan Sprague Dawley, Inc., Indianapolis, IN). In study I we exam-ined the effects of intrarenal NOsynthase inhibition on renal hemody-namics and excretory function when renal perfusion pressure was keptconstant with a suprarenal aortic snare. In study II we examined theeffects of chronic exposure to a high dietary salt intake on endogenousNOactivity.

Study L Role of increased renal perfusion pressure in theeffects of NOsynthase inhibition on renal hemodynamicsand excretory functionClearance studies. Rats were anesthetized (Inactin, 100 mg/kg i.p.)and surgically prepared for determination of intra-arterial pressure,GFR, renal plasma flow (RPF), and urinary excretion of sodium,water and protein, by clearance techniques as previously described ( 14,

15). The left femoral artery was cannulated with a PE-50 catheter forblood pressure monitoring (transducer model TNf; Gould Inc., Medi-cal Products Div., Oxnard, CA) and blood sampling. Renal perfusionpressure (RPP) was assumed to be equivalent to the arterial pressuremeasured in the left femoral artery. A solution of normal saline con-taining [methoxy-3H]inulin (4 uCi/ml) and p-aminohippurate(PAH, 20 mg/ml; Merck Sharpe & Dohme, West Point, PA) was in-fused at a rate of 1.2 ml/h after a priming dose of 0.5 ml. To maintaineuvolemia, 5%bovine serum albumin in normal saline (Sigma Chemi-cal Co., St. Louis, MO)was infused at 1.5 ml/h after a priming dose of1% body weight. Timed urine collections were obtained from the leftureter for gravimetric determination of urine flow rate, with bloodcollected at the midpoint of each clearance period (250 1l).

For renal artery infusions, a PE-10 catheter with a curved atten-uated tip was advanced in a retrograde fashion from the right femoralartery and placed in the orifice of the left renal artery under directvisualization. In this fashion, dissection and manipulation of the renalartery were unnecessary and thus potential denervation or vasculardamage avoided. To maintain patency in the renal artery catheter,heparinized saline was infused at a rate of 0.5 ml/h when other agentswere not being administered.

Analytical techniques. Radioactivity in urine and serum was quan-titated by liquid-scintillation counting. PAHlevels in urine and serumwere determined by AutoAnalyser (Technicon Instruments Corp.,Tarrytown, NY). Urine sodium was determined by flame photometer(Instrumentation Laboratory Inc., Lexington, MA). Urinary total pro-tein excretion was assayed by the Coomassie dye method (Bio-RadLaboratories, Richmond, CA). GFRand RPF were calculated withstandard formulas. Renal blood flow (RBF) was calculated by dividingRPF by 1 - hematocrit and renal vascular resistance was calculated(RVR) by dividing RPPby renal blood flow.

Experimental protocols. Two groups of rats (n = 8 in each group)were studied after overnight fast. Group I rats (control) were allowed toequilibrate for 30 min after surgical preparation and then two baseline( 15 min) clearances were performed. The intrarenal infusion was thenchanged to L-NMMA (generous gift of S. Moncada, Wellcome Re-search Laboratories, Beckenham, Kent, UK) dissolved in normal sa-line and delivered at 250 and 500ug - kg-' per min. There was a 1 5-minequilibration period at each dose level followed by two 1 5-min clear-ance periods. In group II (snare) rats, a length of suture was placedaround the abdominal aorta, proximal to the renal arteries, duringsurgical preparation. The ends of the suture were suspended in an appa-ratus that allowed graded upward tension to be applied. Group II ratsthen underwent a protocol identical to that in group I with the excep-tion that after the baseline clearances, the tension on the aortic snarewas continuously adjusted to maintain RPPat baseline levels. An addi-tional group of rats (n = 6) was prepared for clearance studies in amanner identical to that of group II (snare) except that no tension wasapplied to the aortic snare. Saline vehicle was infused into the renalartery (0.5 ml/h) for the duration of the experiment. Renal functionwas determined according to the exact protocol described above, thusproviding a time-control group.

Study I. Experimental protocolsEffect of salt loading on serum NO2 + NO3 levels and urinary NO2+ NO3excretion rates. Two groups of rats (n = 6 in each group) werestudied. On day 0, while consuming standard rat chow and tap water ablibitum, all rats were placed in metabolic cages and urine was collectedfor 24 h in sterile tubes containing antibiotic/antimycotic solution(100 U/ml penicillin, 100 U/ml streptomycin, and 0.25 yg/ml am-photericin B). Blood was obtained by distal tail-vein sampling fordeter-mination of serum NO2+ NO3. Group III (control) was continued onstandard rat chow and tap water ad libitum. Group IV (salt loaded) wasprovided the same chow, but was switched to 1% NaCl solution as

drinking water. After 2 wk rats were placed in metabolic cages, urinecollected for 24 h, and final blood samples obtained by cardiac punc-ture after brevital anesthesia.

Nitric Oxide and Adaptation to Increased Dietary Salt Intake 643

Serum and urine samples were assayed for the stable NOend prod-ucts, NO2and NO3, as previously described (26-28). Briefly, sampleswere incubated for I h at 370C with Escherichia coli nitrate reductase,prepared from bacteria grown under anaerobic conditions, effectingcomplete reduction of NO3 in the sample to NO2. NO2in the samplewas then quantitated using the Griess reagent, as previously described(25, 26, 29). Known concentrations of NaNO2and NaNO3were usedas standards in each assay and each sample was run in duplicate. TheNO2 measured in this way reflects the sum of NO2 and NO3 in theoriginal sample. Urinary excretion was calculated by multiplying thetotal concentration of NO2 + NO3 in each sample by 24-h urinaryvolume. Because rats exhibited significant increases in body weightover the course of the experiment, NO2+ NO3values were normalizedfor body weight to allow comparison of basal and final values. Baselineand final urine samples were also assayed for cGMPconcentration byradioimmunoassay as previously described (14, 25) and for sodiumconcentration by flame photometer.

Effect of chronic salt loading on renal hemodynamic responses toNOsynthase inhibition. Two additional groups of rats (n = 10 in eachgroup) were studied. Group V (control) was provided standard ratchow and tap water, whereas group VI (salt loaded) received standardrat chow and 1%NaCl in the drinking water, for 2 wk. At the end of thistime period rats in groups V and VI were surgically prepared for clear-ance studies to evaluate renal function before and after intrarenal infu-sion of the NOsynthase inhibitor, L-NAME, infused at 50 and 125

Lg- kg-' per min using a protocol identical to that described above forgroup I rats. Pilot studies demonstrated that renal artery infusion of thehigher dose of L-NAME (125 ug. kg-' per min) results in maximalrenal hemodynamic responses (unpublished observations).

Statistical analysisData are presented as mean±SEM. Differences for parameters betweenbaseline and subsequent time points within one group were comparedby analysis of variance for repeated measures and subsequent Scheffetest. Differences between two groups at a given time point were com-pared by unpaired Student's t test. Differences were considered signifi-cant for P < 0.05. All statistical analyses were done with STATVIEW512 software (Brainpower, Calabases, CA).

ResultsStudy L Role of renal perfusion pressure in responseof renal hemodynamics and excretory functionto NOsynthase inhibitionThere was no difference in body weight (384±7 vs. 390±6 g, P= NS) or left kidney weight (1.39±0.04 vs. 1.41±0.03 g, P

= NS) between group I (control) and group II (aortic snare)rats. Renal hemodynamic parameters in the basal state and inresponse to intrarenal arterial infusion of L-NMMAat 250 and500 ,g kg-' per min are shown in Table I. In control ratsintrarenal arterial infusion of L-NMMAinduced a dose-depen-dent increase in RPP, reflecting a systemic pressor responseoccuring despite intrarenal infusion. Both GFRand RPFpro-gressively and significantly decreased during L-NMMAinfu-sion, in association with marked renal vasoconstriction. Ingroup II (aortic snare) rats initial RPPwas slightly, but signifi-cantly elevated compared with control animals; however, RPPwas effectively maintained at basal levels during L-NMMAin-fusion. There was no difference in GFR, RPF, or RVRbetweengroup I and group II rats in the basal state, indicating thatplacing the aortic snare did not effect basal renal function. Insnared rats GFRand RPFdecreased in a dose-dependent fash-ion in response to intrarenal L-NMMAinfusion, falling to sig-nificantly lower levels as compared with group I at the highestdose. L-NMMAinfusion also resulted in renal vasoconstric-tion, as reflected in a progressive increase in RVR. The magni-tude of the increase in RVRwas similar in group I and group IIrats. Thus, intrarenal EDRFinhibition resulted in exaggerateddecreases in GFRand RPFwhen the concomitant increase inRPPwas prevented.

In time control rats an aortic snare was placed, but no ten-sion was applied, and saline vehicle was infused into the renalartery. Renal hemodynamics were determined at time pointscorresponding to measurements in groups I and II. This groupof rats had body weights (379±10 g) and left kidney weights(1.34±0.04 g) that were similar to those of rats in groups I andII. Hematocrit was stable over the time course of the experi-ment (basal vs. final, 51.2±0.9 vs. 50.5±1.0%, P = NS). RPPdid not change over the course of the experiment in time con-trol rats (basal vs. final, 121±4 vs. 118±2 mmHg, P = NS).GFRhad decreased somewhat by the end of the protocol (basalvs. final, 0.98±0.06 vs. 0.76±0.03 ml/min, P < 0.05), how-ever, RPF(basal vs. final, 5.01±0.29 vs. 5.44±0.36 ml/min, P= NS) and RVR(basal vs. final, 12.09±0.82 vs. 11.05±0.71mmHg.min-1 per ml, P = NS) were unchanged over thecourse of the experiment.

The effect of intrarenal infusion of L-NMMAon urinaryexcretion of sodium, water, and protein in rats of groups I and

Table L Renal Hemodynamic Parameters of Rats in Study I

Group RPP GFR RPF RVR

mmHg mil/min mmHg*min' per ml

Group IControl, baseline 116±3 0.97±.04 5.47±.36 11.04±.65L-NMMA

250,gg-kg-' per min 129±3* 0.86±.02* 4.69±.18* 14.09±.59*500 tg/kg/min 138±2* 0.69±.02* 3.88±.28* 18.83±1.06*

Group IISnare, baseline 126±2t 0.99±.06 5.52±.31 11.39±.56L-NMMA

250 tg/kg/min 126±2 0.85±.02* 4.64±.24* 13.60±.71500 ug/kg/min 126±2t 0.61±.03*t 3.21 ±.24*t 20.69±1.98*

Refer to glossary for definition of terms. All values expressed as mean±SEM. * P < 0.05 vs. baseline; t p < 0.05 group II vs. group I same period.


a Control GoupI l

*-w AorticSnare Group 11|

250 500L-NMMA (gg/kg/mln)

t.

250 500L-NMMA (gg/kg/mln)

was observed for urine flow rate (Fig. 1 B). In group I rats,L-NMMA infusion resulted in a dose-dependent increase inurine flow rate. In contrast, in group II rats, infusion of L-NMMAinduced a marked antidiuretic effect. Fig. 1 Cshowsthe effect of intrarenal L-NMMAinfusion on urinary proteinexcretion rate. As can be seen in group I rats, when RPPisallowed to increase, the highest dose of L-NMMAinduced asignificant increase in protein excretion rate. The proteinuriceffect of intrarenal L-NMMAwas completely prevented whenRPP was maintained at basal levels (group II). Thus, L-NMMA-induced proteinuria was dependent on a concomitantincrease in RPP.

Study II. Effect of salt-loading on serum NO2+ NO3levelsand urinary NO2+ NO3excretion ratesBody weight increased significantly over the 2-wk experimentalperiod in both group III (control, baseline vs. final, 280±3 vs.366±6 g, P < 0.05) and group IV (salt-loaded, 281±2 vs.355±2 g, P < 0.05) rats, however, final body weight was notdifferent between control (group III) and salt-loaded (groupIV) rats. Urine sodium excretion increased slightly in group IIIrats (baseline vs. final, 0.78±0.27 vs. 1.90±0.12 meq/24 h, P< 0.05) consistent with growth. In group IV rats, urine sodiumexcretion increased dramatically in response to salt loading(baseline vs. final, 0.63±0.14 vs. 5.30±0.66 meq/24 h, P< 0.05, P < 0.05 vs. group III).

60

i

z D 50-+ 0

NCI0 0z O4-

= EA2 E 40 -

E0

r-C

30Baseline 250 500

L-NMMA (gg/kg/min) ABaseline 2 wk

Figure 1. Renal excretory function in rats of study I in the baselineand after renal arterial infusion of the NOsynthase inhibitor, L-

NMMA.Urinary excretion rates of sodium (A), water (B), and pro-tein (C) are shown for control rats (group I), in whomRPPvariedspontaneously, and aortic-snare rats (group II), in whomRPPwas

fixed at baseline levels. *P < 0.05 vs. baseline; tP < 0.05 group I vs.

group II.

II is shown in Fig. 1. In control rats, in whomRPPwas allowedto increase, a significant increase in sodium excretion com-

pared with baseline was observed at both doses of L-NMMA(Fig.. 1 A). In contrast, in snared rats, in whomRPPremainedat basal levels, urinary sodium excretion did not increase at 250,gg - kg-' per min and significantly decreased compared withbaseline at the higher dose of L-NMMA. In snared rats, urinarysodium excretion was significantly decreased compared withcontrol rats at both dose levels of L-NMMA. Thus, when theincrease in RPP is prevented, intrarenal NOsynthase inhibi-tion has a significant antinatriuretic effect. A similar pattern

Control

*1111 R<aft I nadpef

Group III

tramlin IV

83:z

NO

0 0 4000Z M

C0

°: 3000

X

0

2000. C

1000

B Baseline 2 wk

Figure 2. Serum concentrations of N02 + N03 (A) and urinary N02+ N03 excretion rate (B) in rats of study II, in the baseline and after2 wk of either normal (group III) or increased (group IV) dietary saltintake. *P < 0.05 vs. baseline; tP < 0.05 group III vs. group IV.


5.

2 4-

LUc 3

E

U) -2

11

ABaseline

30 -

20 -

10 I

C

E

0

IC0

BaselineB

4)6

a

C

0

p

'C

0

a.ES.

a.

C1

30

20 -

Serum levels of NO2+ NO3in group III and group IV ratsat baseline and after the 2-wk experimental period are shown inFig. 2. Baseline values for serum NO2+ NO3were not differentbetween the two groups (Fig. 2 A). After 2 wk the serum NO2+ NO3was 41.1±+3.7 nmol * ml-' per 100 g body wt in controlrats, a value not significantly different from baseline in thisgroup. In contrast, the serum levels of NO2+ NO3rose signifi-cantly in group IV (salt-loaded) rats to 51.9±3.6 nmol. ml-'per 100 g body wt. This value was significantly higher thanbaseline for this group (P < 0.05) and, in addition, was signifi-cantly higher than the 2-wk value in group II (control, P< 0.05).

Fig. 2 B demonstrates a significant rise in urinary excretionof NO2+ NO3 in both the control (group III) and salt-loaded(group IV) rats between baseline and 2 wk (Fig. 2 B). How-ever, at the end of the 2-wk experimental period, urinary excre-tion of NO2_+ NO3was significantly higher in salt-loaded ascompared with control rats (group IV vs. group III, 4,078±478nmol * 24 h-' per 100 g body wt vs. 3,123±182 nmol - 24 h-'per 100 gmbody wt, P< 0.05). Thus, in these normal rats, saltloading was associated with a significant increase in the serumconcentration and urinary excretion of NO2+ NO3. Further-more, as shown in Fig. 3, when baseline and final values of ratsin groups III and IV were considered, there was a highly signifi-cant correlation between urinary excretion of NO2+ NO3andurinary sodium excretion (R2 = 0.728, P < 0.001).

Group III and group IV rats had similar urinary cGMPexcretion rates at baseline (Fig. 4), and cGMPexcretion rateincreased significantly compared with baseline after 2 wk inboth groups of rats. However, as shown in Fig. 4, salt-loadedrats (group IV) demonstrated a greater increase in urinarycGMPexcretion rate, reaching a final level that was signifi-cantly increased as compared with control rats (group III).Similar to the urinary NO2 + NO3 excretion, urinary cGMPexcretion was also significantly correlated with urinary sodiumexcretion (P < 0.001 ). Finally, we also found a significant andpositive correlation between urinary excretion of NO2+ NO3and cGMP(P < 0.01, data not shown), thus confirming anassociation between these three urinary parameters, andstrongly suggesting that salt loading causes increased produc-tion of endogenous NO.

0

-

co

._.

m

x .=co0)Z00

z EcJ C

h..

0 -

C._) zt

17500

7500

2500

S

0 Control Group 11

E 16 | *-SaltLoaded Group IV

0

O 12

4

0

co

0.

Baseline 2 wk

Figure 4. Effect of dietary salt intake on urinary excretion of cGMPinrats of study II. Urinary cGMPexcretion is shown in the baselineand after 2 wk on normal (group III) or increased (group IV) salt in-take. * P < 0.01I vs. baseline; t P < 0.05 group III vs. group IV.

Effect of chronic salt loading on renal hemnodynamicresponses to NOsynthase inhibitionValues for body weight, left kidney weight, 24-h urine volume,and sodium excretion for rats in group V (control) and groupVI ( salt loaded ) are shown in Table II. As can be seen, althoughhigh dietary salt intake did not significantly effect body weight,salt loading was associated with a significant increase in kidneyweight. As expected, urinary sodium excretion rate increasedmarkedly in response to increased salt intake in group VI rats.Urine volume tended to increase as well in group VI- comparedwith group V rats, but this difference did not achieve statistical

x *r

signifcance.Renal hemodynamic responses to intrarenal NOsynthase

inhibition are shown in Table III. L-NAME infusion induced adose-dependent increase in RPPthat was similar in magnitudein control and salt-loaded rats. As expected, group V ( control )rats demonstrated a dose-dependent decrease in GFR andRPF, whereas RVR significantly increased. These hemody-namic effects were similar to those observed in study I afterrenal artery infusion of L-NMMA. In group VI rats thesechanges were directionally similar; however, at the highest doseof L-NAME, GFRfell to a significantly lower level as comparedwith group V rats. The fall in RPFand increase in RVRalsotended to be of greater magnitude in salt-loaded as comparedwith control rats, but these differences did not achieve statisti-cal significance. To facilitate comparison of the magnitude of

2 Table 2. Characteristics of Rats in Group V (Control)R = 0.728 and Group VI (Salt loaded) after 2-wk Experimental PeriodP < 0.001

0 2000 4000 6000Urinary Na Excretion Rate

(geq/24 h)

8000

Figure 3. Regression line of urinary NO2 + NO3excretion rate onurinary sodium excretion rate in rats of study II. Data points are forrats in baseline condition and after 2 wk on either normal (groupIII) or increased (group IV) salt intake.

Left Urine Urine sodiumGroup Body weight kidney weight volume excretion

g ml/24 h meq/24 h

Group V,control 391±4 1.45±.03 16.2±2.2 2.04±.37

P value NS <0.01 NS <0.001Group VI,

salt loaded 406±9 1.65±.05 20.6±4.6 8.10±1.10


Table III. Renal hemodynamic parameters of rats in Study II

Group RPP GFR RPF RVR

mmHg mil/min mmHgmin' per ml

Group VControl, baseline 121±4 0.93±.05 5.64±.41 11.09±.83L-NAME

50,g -kg-' per min 133±4* 0.73±0.4* 3.33±.28* 21.83±2.47*125 ,g/kg/min 146±3* 0.67±.03* 3.00±.23* 26.18±2.42*

Group VISalt-loaded, baseline 124±4 1.02±.06 5.79±.25 10.85±.57L-NAME

50,gg/kg/min 135+4* 0.73±.04* 3.31±.17* 20.81±1.25*125 jug/kg/min 145±4* 0.60±.03*t 2.57±.17* 30.00±2.69*

Refer to glossary for definition of terms. All values expressed as mean±SEM.* P < 0.05 vs. baseline; tp < 0.05 group VI vs. group V same period.

the hemodynamic responses of groups V and VI to the highestdose of L-NAME, Fig. 5 shows these data expressed as percentchange from baseline values. As can be seen in Fig. 5 A, thesystemic pressor response was not different in control and salt-loaded rats, however, the renal vasoconstrictor response tohigh-dose L-NAME tended to be greater in group VI (salt-loaded) as compared with group V (control) rats (P = 0.06).The decrease in GFRand RPF, expressed as percentage changefrom baseline values, was significantly greater in salt-loaded ascompared with control rats (Fig. 5 B). Thus, in rats exposed to

200

a

0

a

E0

0,cmSa

RPP RVRA

10ISc

0

am

0

0ICU

C)

-10-

-20 -

-30 -

-40 -

-50 -

lV:(f",P<

1~~~~~~~~I -. I4~

RPFB

Figure 5. Effect of renal arterial infusion of the NOsynthase inhibitor,L-NAME (125 Ag- kg-' per min), on renal hemodynamics in rats ofstudy II, after 2 wk on normal (group V) or increased (group VI) di-etary salt intake. (A) Percent change from baseline in RPPand RVR.(B) Percent change from baseline in GFRand RPF.

a high dietary salt intake for 2 wk, there was a greater renalhemodynamic response to maximal NOsynthase inhibitionthan was observed in control rats.

Discussion

Over the last decade it has become increasingly apparent that,in addition to regulation of vascular permeability and controlof hemostasis, the vascular endothelium is important in theregulation of vascular tone and regional hemodynamics (30).Investigation in this area has made it clear that control of vascu-lar function by the endothelium is quite complex and involvesthe balanced synthesis and release of both vasodilator (prosta-cyclin, EDRF/NO) and vasoconstrictor (endothelin, endothe-lium-derived contracting factor) substances (31). Modulationof regional hemodynamics by the endothelium has perhapsbeen best demonstrated by investigations into renal hemody-namics and excretory function.

It has become clear that the vasodilatory action of acetyl-choline, when studied in vitro, is due to stimulation and releaseof the EDRF, NO( 1-3). Tolins and coworkers ( 14) recentlydemonstrated that the renal hemodynamic effects of acetylcho-line in vivo are also mediated by the EDRF, NO. In experimen-tal animals acetylcholine infusion results in decreased systemicblood pressure, a large increase in glomerular capillary plasmaflow, but no change in single-nephron GFR, resulting in a fallin filtration fraction (32, 33). Acetylcholine infusion has alsobeen reported to have a diuretic and natriuretic effect (33, 34).Lahera et al. (35) reported that this acetylcholine-induced di-uretic effect was dependent on intact NOsynthase activity.Thus, in vivo experimental evidence suggests that stimulationof NOsynthesis results not only in vasodilatation, but diuresisand natriuresis as well.

Further evidence supporting a role of NOin modulation ofrenal function has come from studies using inhibitors of NOsynthase in vivo ( 15-20). These studies suggest that inhibitionof endogenous NO synthase results in decreased GFRandrenal vasoconstriction with a subsequent increase in filtrationfraction. On a microcirculatory level, these changes are asso-ciated with a predominant vasoconstrictor effect on the post-glomerular circulation, an increase in glomerular capillary


TF

CFRl

El Group V

El Group VI

pressure, and a fall in glomerular ultrafiltration coefficient.However, interpretation of these studies has been complicateddue to the simultaneous systemic effects of these NOsynthaseinhibitors, resulting in increased RPP. It has remained unclearto what degree the effects of inhibition of NOsynthase on renalhemodynamics (increased RVR) and excretory function (di-uresis and natriuresis) are direct or due to renal responses toincreased RPP, such as autoregulation (36) and pressure natri-uresis (37), respectively.

In this regard, Radermacher et al. (38) observed that bothL-NMMAand nitro-L-arginine increased renal vascular resis-tance in the isolated perfused kidney of the rat, an experimen-tal model free from systemic hemodynamic influences. Laheraet al. ( 18) investigated the renal functional response to intrave-nous infusions of L-NAME in the rat. They reported that dur-ing low-dose L-NAME infusion (1 igg kg-' per min i.v.),which did not significantly change systemic arterial pressure,there was a significant decrease in renal sodium and water ex-cretion rate. In contrast, higher doses of L-NAME(50 ,g . kg-'per min) induced an increase in systemic blood pressure andwere associated with a diuretic and natriuretic effect. Boththese studies suggested that NOsynthase inhibition had effectson renal hemodynamics and excretory function that were inde-pendent of systemic hemodynamic responses.

Study I was designed as an attempt to discern direct fromindirect effects of NOsynthase inhibition on both renal hemo-dynamics and excretory function. Our experimental model al-lowed evaluation of responses of the intact in situ kidney tomaximal intrarenal NOsynthase inhibition, while preventingany influence of alterations in RPP. Wehave demonstratedthat the renal vasoconstriction and fall in GFRobserved duringNOsynthase inhibition occur, and are perhaps accentuated,when renal perfusion pressure remains constant. This suggeststhat the renal hemodynamic effects of NOsynthase inhibitionare due to a direct action on the kidney and not secondary toadaptations such as renal autoregulation. In contrast, it is ap-parent that the diuresis and natriuresis observed during NOsynthase inhibition are indeed secondary to and dependent onthe concomitant increase in RPP, and thus represent a pressurenatriuresis. These results are in agreement with the work ofJohnson and Freeman (2 1), who previously suggested that sys-temic inhibition of NOsynthase produces a pressure-depen-dent natriuresis. The results of study I are also consistent withprevious reports of a natriuretic effect of agonists of the EDRF,NO, such as acetylcholine (33, 34); as well as the work ofAlberola et al. (39) in which it was reported that, in the dog, thenatriuresis and diuresis induced by extracellular volume ex-pansion with saline was blunted during intrarenal L-NAMEinfusion. These previous studies suggest that stimulation of NOactivity increases renal sodium excretion. Our data suggeststhat the direct renal effect of inhibition of NO activity is adecrease in sodium excretion. Taken together, these resultssuggest that activity of the endogenous NOsystem may play animportant role in modulation of renal sodium handling.

In study I we also made the novel observation that acuteNOsynthase inhibition has a significant proteinuric effect andthat this increase in protein excretion is completely preventedby fixing renal perfusion pressure at basal levels. This suggeststhat the proteinuric effect of NOsynthase inhibition is second-ary to acute hemodynamic alterations in association with anincrease in RPP, most likely increased glomerular capillary

pressure and increased filtration fraction (40), hemodynamicchanges previously observed during NOsynthase inhibition byZatz and DeNucci (16). Because the increase in proteinuriawas prevented in snared rats, we can speculate that the increasein glomerular capillary pressure observed after NOsynthaseinhibition ( 16) also depends on an increase in RPPand trans-mission of this pressure to the glomerular capillary bed. It isinteresting that identical glomerular hemodynamic effects andproteinuria have been demonstrated during exogenous angio-tensin II infusion in the rat (41 ). Indeed, previous reports havesuggested that part of the renal hemodynamic effect of NOsynthase inhibition may be secondary to increased activity ofangiotensin II within the glomerular circulation ( 15, 20). Itshould be emphasized that the mechanism underlying L-NMMA-induced proteinuria was not directly addressed in thecurrent study and so the importance of glomerular microcircu-latory effects must remain speculative.

The results of study I suggested that the endogenous NOsystem may play a role in renal sodium handling, with in-creased NO activity stimulating sodium excretion and de-creased NOactivity resulting in sodium retention. Further-more, a recent report by Chen and Sanders (22) suggested thata defect in NOsynthesis may underlie the phenomenon ofsalt-sensitive hypertension in rats. These investigators notedthat, in Dahl rats, an increased dietary salt load resulted inincreased activity of the NOsystem in salt-resistant, but notsalt-sensitive rats. They hypothesized that the development ofsalt-sensitive hypertension may be secondary to an inability toenhance NO synthesis in response to increased dietary saltload. However, the conclusions of this study were based pre-dominantly on the hemodynamic response to NOsynthaseinhibitors or the response to exogenous L-arginine, rather thandirect measurements of NOmetabolites. Furthermore, it is pos-sible that the defect in NOproduction in hypertensive Dahlrats is a result rather than a cause of salt-sensitive hypertension(42). Study II was designed to further define the relationshipbetween the adaptation to increased dietary salt intake and theendogenous NOsystem in normal rats.

In study II we observed that the serum concentration andurinary excretion of the NOdecomposition products, NO2+ NO3, were increased in salt-loaded rats as compared withcontrol rats, providing evidence that NOproduction increasesin response to increased dietary salt intake in normal rats. Al-though the circulating and excreted NO2 + NO3may not beexclusively derived from metabolism of NO, our previous stud-ies in rats given endotoxin (26) as well as a recent study using"5N-labeled substrate for NOsynthase in humans (43) haveestablished that changes in serum and urine levels of NO2+ NO3do reflect in vivo production of NO. The source of theincreased NOdetected in response to increased dietary saltintake cannot be determined from the present studies. How-ever, ex vivo incubations of tissues from rats exposed to endo-toxin suggest that both vascular tissue and the kidney, particu-larly the glomerulus, are a major source of inducible NOpro-duction (26, 44). Whether this holds true in the setting ofsalt-loading will need to be determined in future studies.

The conclusion that NOproduction increases in responseto increased dietary salt intake is supported by the finding of asignificant correlation between urinary sodium excretion, a re-flection of intake, and urinary NO2+ NO3excretion. Similarly,urinary sodium excretion also correlated with urinary cGMP


excretion, another biological marker of NOactivity in vivo( 14, 26). Moreover, there was a significant correlation betweenurinary NO2+ NO3and cGMPexcretion rates, providing fur-ther evidence that these two markers are related, most likely byreflecting endogenous NOactivity. The urinary cGMPexcre-tion rate was significantly increased in salt-loaded as comparedwith control rats. This latter finding must be interpretedcautiously, however, as an increase in atrial natriuretic peptidelevels in response to salt loading could be contributory (44).

The hemodynamic responses of rats in study II to NOsyn-thase inhibition also support our conclusion that increased di-etary salt intake increases endogenous NOactivity. After 2 wkof exposure to a high dietary salt intake, although baseline renalhemodynamic parameters were similar, intrarenal infusion ofL-NAME was associated with a greater decrease in GFRandRPFand a tendency to an enhanced RVRresponse comparedwith control rats. This suggests that the basal renal hemody-namic state was characterized by enhanced endogenous NOactivity in the salt-loaded group.

In summary, the current study demonstrates that inhibi-tion of NOsynthase in vivo results in direct renal hemody-namic and excretory effects, independent of changes in RPP.In particular, NOsynthase inhibition is associated with a de-crease in urinary sodium excretion when a concomitant in-crease in RPP is prevented. Furthermore, exposure of normalrats to a high salt diet for 2 wk induces increased NOproduc-tion as determined by increases in serum concentrations ofNO2+ NO3and increased urinary excretion rate of NO2+ NO3and cGMP. Chronically salt-loaded rats also demonstrated anenhanced renal hemodynamic response to NOsynthase inhibi-tion compared with rats with a normal salt intake. Wecon-clude that the endogenous NOsystem participates in the physi-ological renal adaptation to increased dietary salt intake. Clari-fication of the role of NOin salt-induced hypertension andother pathological conditions associated with abnormal so-dium homeostasis awaits future studies.

Acknowledaments

Wethank B. Nyguyen, K. Lewis, and K. Anderson for technical assis-tance and D. Caruso and M. Helgerson for assistance with manuscriptpreparation.

This study was supported by funds from the Department of Vet-erans Affairs.

References

1. Furchgott, R. F., and V. Zawadski. 1980. The obligatory role of endothelialcells in the relaxation of arterial smooth muscle by acetylcholine. Nature (Lond.).299:373-376.

2. Ignarro, L. J., G. M. Buga, K. S. Wood, R. E. Byrns, and G. Chandhuri.1987. Endothelium-derived relaxing factor produced and released from arteryand vein is nitric oxide. Proc. Natl. Acad. Sci. USA. 84:9265-9269.

3. Amezcua, J. L., G. J. Dusting, R. M. J. Palmer, and S. Moncada. 1988.Acetylcholine induces vasodilatation in the rabbit isolated heart through releaseof nitric oxide, the endogenous nitrovasodilator. Br. J. Pharmacol. 95:830-834.

4. Hutchinson, P. J. A., R. M. J. Palmer, and S. Moncada. 1987. Comparativepharmacology of EDRFand nitric oxide on vascular strips. Eur. J. Pharmacol.141:445-451.

5. Palmer, R. M. J., D. S. Ashton, and S. Moncada. 1988. Vascular endothe-lial cells synthesize nitric oxide from L-arginine. Nature (Lond.). 333:664-666.

6. Palmer, R. M. J., D. D. Rees, D. S. Ashton, and S. Moncada. 1988. L-argin-

ine is the physiologic precursor for the formation of nitric oxide in endotheliumdependent relaxation. Biochem. Biophys. Res. Commun. 153:1251-1256.

7. Palmer, R. M. J., and S. Moncada. 1989. A novel citrulline-forming en-zyme implicated in the formation of nitric oxide by vascular endothelial cells.Biochem. Biophys. Res. Commun. 158:348-352.

8. Bredt, D. S., and S. H. Snyder. 1990. Isolation of nitric oxide synthetase, acalmodulin-requiring enzyme. Proc. Nall. Acad. Sci. USA. 87:682-685.

9. Rees, D. D., R. M. J. Palmer, H. F. Hodson, and S. Moncada. 1989. Aspecific inhibitor of nitric oxide formation from L-arginine attenuates endothe-lium-dependent relaxation. Br. J. Pharmacol. 96:418-424.

10. Dubbin, P. N., M. Zambetis, and G. J. Dusting. 1990. Inhibition of endo-thelial nitric oxide biosynthesis by N-nitro-L-arginine. Clin. Exp. Pharmacol.Physiol. 17:281-286.

11. Moore, P. K., 0. A. al-Swayeh, N. W. S. Chong, R. A. Evans, and A.Gibson. 1990. L-N0-nitro arginine (L-NOARG), a novel, L-arginine-reversibleinhibitor of endothelium-dependent vasodilation in vitro. Br. J. Pharmacol.99:408-412.

12. Rappaport, R. M., and F. Murad. 1983. Agonist-induced endothelium-de-pendent relaxation in rat thoracic aorta may be mediated through cGMP. Circ.Res. 52:352-357.

13. Shultz, P. J., A. E. Schorer, and L. Raij. 1990. Effects of endothelium-de-rived relaxing factor and nitric oxide on rat mesangial cells. Am. J. Physiol.258:F1 62-F167.

14. Tolins, J. P., R. M. J. Palmer, S. Moncada, and L. Raij. 1990. Role ofendothelium-derived relaxing factor in regulation of renal hemodynamic re-sponses. Am. J. Physiol. 258:H655-H662.

15. Tolins, J. P., and L. Raij. 1991. Effects of amino acid infusion on renalhemodynamics. Role of endothelium-derived relaxing factor. Hypertension (Dal-las). 17:1045-1051.

16. Zatz, R., and G. DeNucci. 1991. Effects of acute nitric oxide inhibition onrat glomerular microcirculation. Am. J. Physiol. 261:F360-F363.

17. Baylis, C., P. Harton, and K. Engels. 1990. Endothelial derived relaxingfactor controls renal hemodynamics in the normal rat kidney. J. Am. Soc.Nephrol. 1:875-881.

18. Lahera, V., M. G. Salom, F. Miranda-Guardiola, S. Moncada, and J. C.Romero. 1991. Effects of N0-nitro-L-arginine methyl ester on renal function andblood pressure. Am. J. Physiol. 261:F1033-F1037.

19. King, A. J., J. L. Troy, S. Anderson, J. R. Neuringer, M. Gunning, andB. M. Brenner. 1991. Nitric oxide: a potential mediator of amino acid-inducedrenal hyperemia and hyperfiltration. J. Am. Soc. Nephrol. 1:1271-1277.

20. DeNicola, L., R. C. Blantz, and F. B. Gabbai. 1992. Nitric oxide andangiotensin II. Glomerular and tubular interaction in the rat. J. Clin. Invest.89:1248-1256.

21. Johnson, R. A., and R. H. Freeman. 1992. Pressure natriuresis in ratsduring blockade of the L-arginine/nitric oxide pathway. Hypertension (Dallas).19:333-338.

22. Chen, P. Y., and P. W. Sanders. 1991. L-Arginine abrogates salt-sensitivehypertension in Dahl/Rapp rats. J. Clin. Invest. 88:1559-1567.

23. Marletta, M. A., P. S. Yoon, R. Lyengar, C. D. Leaf, and J. S. Wishnok.1988. Macrophage oxidation of L-arginine to nitrite and nitrate: nitric oxide is anintermediate. Biochemistry. 27:8706-871 1.

24. Beasley, D. J., J. H. Schwartz, and B. M. Brenner. 1991. Interleukin 1induces prolonged L-arginine-dependent cyclic guanosine monophosphate andnitrite production in rat vascular smooth muscle cells. J. Clin. Invest. 87:602-608.

25. Shultz, P. J., M. A. Tayeh, M. A. Marletta, and L. Raij. 1991. Synthesisand action of nitric oxide in rat glomerular mesangial cells. Am. J. Physiol.26 1:F600-F606.

26. Shultz, P. J., and L. Raij. 1991. Endogenously synthesized nitric oxideprevents endotoxin-induced glomerular thrombosis. J. Clin. Invest. 90:1718-1725.

27. Bartholomew, B. 1984. A rapid method for the assay of nitrate in urineusing the nitrate reductase enzyme of Escherichia coli. Ed Chem. Toxicol.22:541-543.

28. Granger, D. L., J. B. Hibbs, Jr., J. R. Perfect, and D. T. Durack. 1990.Metabolic fate of L-arginine in relation to microbiostatic capability of murinemacrophages. J. Clin. Invest. 85:264-273.

29. Green, L. C., D. A. Wagner, J. Glogowski, P. L. Skipper, J. S. Wishnok,and S. R. Tannenbaum. 1982. Analysis of nitrate, nitrite and 5N nitrate in biolog-ical fluids. Anal. Biochem. 126:131-138.

30. Tolins, J. P., P. J. Shultz, and L. Raij. 1991. Role of endothelium-derivedrelaxing factor in regulation of vasculartone and remodeling. Update on humoralregulation of vascular tone. Hypertension (Dallas). 17:909-916.

31. Luscher, T. 1990. The endothelium, target and promoter of hypertension?Hypertension (Dallas). 15:482-485.

32. Baylis, C. W., W. M. Dean, B. D. Myers, and B. M. Brenner. 1976. Effectsof some vasodilator drugs on transcapillary fluid exchange in renal cortex. Am. J.Physiol. 230:1148-1158.


33. Baer, P. G., L. G. Navar, and A. C. Guyton. 1970. Renal autoregulation,filtration rate and electrolyte excretion during vasodilatation. Am. J. Physiol.2 19:6 19-625.

34. Early, L. E., and R. M. Friedler. 1965. Studies on the mechanism ofnatriuresis accompanying increased renal blood flow and its role in the renalresponse to extracellular volume expansion. J. Clin. Invest. 44:1857-1865.

35. Lahera, V., M. G. Salom, M. J. Fiksen-Olsen, L. Raij, and J. C. Romero.1990. Effects of NG-monomethyl-L-arginine and L-arginine on acetylcholinerenal response. Hypertension (Dallas). 15:659-663.

36. Dewan, S. A. M., and L. G. Navar. 1992. Suppression of blood flowautoregulation plateau during nitric oxide blockade in canine kidney. Am. J.Physiol. 262:F40-F46.

37. Romero, J. C., M. D. Bentley, P. M. Vanhoutte, and F. G. Knox. 1989.Intrarenal mechanisms that regulate sodium excretion in relationship to changesin blood pressure. Mayo Clin. Proc. 64:1406-1424.

38. Radermacher, J., U. Forstermann, and J. C. Frolich. 1990. Endothelium-derived relaxing factor influences renal vascular resistance. Am. J. Physiol.259:F9-F17.

39. Alberola, A., J. M. Pinilla, T. Quesada, J. C. Romero, M. G. Salom, andF. J. Salazar. 1992. Role of nitric oxide in mediating renal response to volumeexpansion. Hypertension (Dallas). 19:780-784.

40. Deen, W. M., M. P. Bohrer, and B. M. Brenner. 1979. Macromoleculetransport across glomerular capillaries. Kidney Int. 16:353-365.

41. Bohrer, M. P., W. M. Deen, C. R. Robertson, and B. M. Brenner. 1977.Mechanism of angiotensin II-induced proteinuria in the rat. Am. J. Physiol.233:F13-F21.

42. Luscher, T. F., L. Raij, and P. M. Vanhoutte. 1987. Endothelium-depen-dent responses in normotensive and hypertensive Dahl rats. Hypertension (Dal-las). 9:157-163.

43. Hibbs, J. B., C. Westenfelder, R. Taintor, Z. Vavrin, C. Kablitz, R. L.Baranowski, J. H. Ward, R. L. Menlove, M. P. McMurray, J. P. Kushner, and etal. 1992. Evidence forcytokine-inducible nitric oxide synthesis from L-arginine inpatients receiving interleukin-2 therapy. J. Clin. Invest. 89:867-877.

44. Geryer, R., H. Witzgall, J. Trembley, J. Gutkowska, and P. Hamet. 1985.Rapid increase in plasma and urinary cyclic GMPafter bolus injection of atrialnatriuretic factor in man. J. Clin. Endocrinol. Metab. 61:1217-1219.


adaptation to increased dietary salt intake in...

Documents